This is a three-week project. During the first two weeks you will spend a few minutes each day gathering weather forecast information from the web. If you are working in a group, each group member will be responsible for gathering information for one or two forecasters. You will analyze your data and prepare your report during the third week.
- You will start by finding and copying down websites for the forecasters you will be collecting from. These forecasters could be the local NBC, FOX, ABC, CBS stations, the local Newspapers, The Weather Channel, Weather Underground (National Weather Service), or other potential forecasters. See the Information Sources page for suggestions and examples.
On Monday through Friday of the first two weeks, you will record several types of high temperature forecasts made by each forecaster. You will record four types of forecasts: 1-day, 2-day, 3-day and 4-day. The chart below explains which types of forecasts you will record on each day.
Types of forecasts to record on specific days:1-day forecast2-day forecast3-day forecast4-day forecastMondayforecast for Tuesdayforecast for Wednesdayforecast for Thursdayforecast for FridayTuesdayforecast for Wednesdayforecast for Thursdayforecast for Fridayforecast for SaturdayWednesdayforecast for Thursdayforecast for Fridayforecast for SaturdayThursdayforecast for Fridayforecast for SaturdayFridayforecast for Saturday
The table shows that on Monday, you will record all four types of forecasts. On Friday, you will record only 1-day forecasts. Note that you will never record forecasts for Sunday. Also, you will collect the most data for 1-day forecasts, the second-most data for 2-day forecasts, and so on. [Some sample graphs showing an analysis of Milwaukee, Wisconsin high temperature forecasts, are shown at the bottom of the ‘Guidance‘ page.]
Print out a copy of the data entry worksheet. You can use this worksheet to record your forecasts. The first picture below shows a part of the worksheet.
If you have Microsoft Excel spreadsheet software on your computer, (click here) to get the entire worksheet. If you don’t have Microsoft Excel installed on your computer, you can display and print the worksheet in four parts. Click on the links below to display each part of the worksheet.
- For each forecaster, find the average relative forecast error for all types of high temperature forecasts: 1-day, 2-day, 3-day and 4-day. (The average relative forecast error gives you the average error, including whether the forecasts are too high or too low. The term ‘average’ is the same as the term ‘mean’.) You’ll also find the average absolute forecast error. (The average absolute forecast error gives you the average error, regardless of whether the forecasts are too high or too low.) Procedures for finding these errors are explained on the ‘Guidance‘ page.
- Compare your final results with those in your group.
- Create two line graphs: one showing the average relative error vs. forecast type (1-day, 2-day, 3-day or 4-day), and the other showing the average absolute error vs. forecast type. On each graph, use a different line for each forecaster. [Sample graphs are shown on the ‘Guidance‘ page.]
Prepare individual and group reports
The individual report should list the forecaster(s) for which you collected data, the web sites used, the forecasts themselves, the verification, the average relative and absolute errors for each forecaster and for each forecast type, and a paragraph evaluating each forecaster’s ability to accurately predict high temperatures for one, two, three and four days into the future. The paragraph should also include the range of errors found. The individual report should also include the two graphs described above.
The group report should include a table listing, for each forecaster, the web sites used, the average relative and absolute errors over time along with the complete line graphs showing results for all forecasters. Included with the graphs should be a written paragraph explaining the results of the line graphs along with your groups explanation and conclusion based on the results.